INDEX

Explanations

words related to connections, relationships, and interactions

oai_token-act-pair · gpt-3.5-turbo

references to associations between concepts or variables

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

vae

-0.68

stall

-0.68

sis

-0.65

cale

-0.63

gm

-0.62

 Penal

-0.61

raq

-0.61

rimp

-0.61

 Floating

-0.59

=-=-=-=-=-=-=-=-

-0.59

POSITIVE LOGITS

 associations

1.10

 association

1.01

eer

0.91

ually

0.89

ality

0.88

anguage

0.83

ociation

0.83

eers

0.81

eering

0.81

 affili

0.80

Activations Density 0.011%

words related to connections, relationships, and interactions

references to associations between concepts or variables

No Comments

No Known Activations