INDEX

Explanations

political figures, particularly presidents and prime ministers

oai_token-act-pair · gpt-3.5-turbo

references to political leaders and their actions or statements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 pear

-0.70

 interpol

-0.68

 Californ

-0.66

 grad

-0.66

 Bronx

-0.66

 phantom

-0.65

 dots

-0.65

 feral

-0.64

 Ghosts

-0.64

 soph

-0.64

POSITIVE LOGITS

 Jinping

0.95

imar

0.90

oÄŁan

0.82

vernment

0.80

llor

0.80

uty

0.80

 appoint

0.76

luck

0.75

 vowed

0.74

 appointed

0.73

Activations Density 0.359%

political figures, particularly presidents and prime ministers

references to political leaders and their actions or statements

No Comments

No Known Activations