INDEX

Explanations

words related to specific entities or proper nouns, such as countries, organizations, and technologies

oai_token-act-pair · gpt-3.5-turbo

specific locations, entities, and notable terms related to various subjects

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

planet

-0.59

cause

-0.58

etheless

-0.57

inces

-0.55

rolet

-0.55

Cause

-0.55

————————————————

-0.54

apult

-0.54

kick

-0.53

ragon

-0.52

POSITIVE LOGITS

 meanwhile

0.94

 there

0.91

we

0.86

it

0.83

 however

0.82

 they

0.81

 alone

0.75

0.73

 jargon

0.72

you

0.67

Activations Density 0.452%

words related to specific entities or proper nouns, such as countries, organizations, and technologies

specific locations, entities, and notable terms related to various subjects

No Comments

No Known Activations