INDEX

Explanations

words related to technical or scientific terms and increases the activation for specific words like 'al', 'iology', 'phase', 'max', 'code', and 'hologram'

oai_token-act-pair · gpt-3.5-turbo

technical jargon and terminology related to programming and coding

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ãĥ´ãĤ¡

-0.66

liest

-0.65

Kes

-0.61

 Samar

-0.60

dirty

-0.60

bound

-0.59

responsible

-0.58

ependent

-0.58

qualified

-0.56

Ago

-0.55

POSITIVE LOGITS

unia

0.73

acters

0.72

uli

0.63

ebted

0.61

ught

0.60

puff

0.60

Reply

0.60

Redditor

0.60

 configurations

0.59

 artery

0.59

Activations Density 0.538%

words related to technical or scientific terms and increases the activation for specific words like 'al', 'iology', 'phase', 'max', 'code', and 'hologram'

technical jargon and terminology related to programming and coding

No Comments

No Known Activations