INDEX

Explanations

words related to politics and government

oai_token-act-pair · gpt-3.5-turbo

references to political topics and issues

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

oken

-0.77

VIDEO

-0.62

aque

-0.60

Sent

-0.59

oute

-0.58

 Solitaire

-0.58

Raven

-0.57

 Crescent

-0.57

ŃĶ

-0.56

rique

-0.56

POSITIVE LOGITS

 forb

0.69

 enforced

0.64

 muster

0.63

Enlarge

0.61

 outlaw

0.61

 looms

0.60

 mandated

0.60

 levied

0.60

 enforcement

0.59

 cens

0.59

Activations Density 0.935%

words related to politics and government

references to political topics and issues

No Comments

No Known Activations