INDEX

Explanations

words related to specific occupations or titles

oai_token-act-pair · gpt-3.5-turbo

references to various professions or identities of individuals

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

mbuds

-0.77

 Helpful

-0.66

ecause

-0.64

nih

-0.64

 sshd

-0.61

noon

-0.61

veyard

-0.60

 Liter

-0.59

actionDate

-0.59

 Legislation

-0.59

POSITIVE LOGITS

 underwent

1.05

 admits

0.97

 endured

0.96

 insists

0.96

 believes

0.93

 wore

0.91

 joked

0.91

 opted

0.90

 extraord

0.89

 overcame

0.89

Activations Density 0.135%

words related to specific occupations or titles

references to various professions or identities of individuals

No Comments

No Known Activations