INDEX

Explanations

proper nouns related to government agencies and law enforcement

oai_token-act-pair · gpt-3.5-turbo

repetitive mentions of the word "the."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

thood

-0.80

galitarian

-0.78

ciation

-0.75

iour

-0.75

verning

-0.73

Beta

-0.73

 thereof

-0.73

Ò

-0.73

adeon

-0.72

cil

-0.72

POSITIVE LOGITS

 revelations

1.19

 crackdown

1.15

 feds

1.15

 secretive

1.12

 thefts

1.11

 disclosures

1.10

FBI

1.09

 trove

1.05

 raids

1.04

 arrests

1.03

Activations Density 0.321%

proper nouns related to government agencies and law enforcement

repetitive mentions of the word "the."

No Comments

No Known Activations