INDEX

Explanations

phrases related to political or military backing or support

oai_token-act-pair · gpt-3.5-turbo

references to entities or groups supported by a governing body or organization

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

istics

-0.79

itta

-0.79

Osc

-0.75

 Gutenberg

-0.74

ijn

-0.73

ogens

-0.73

ennes

-0.73

ciation

-0.70

etics

-0.69

thora

-0.69

POSITIVE LOGITS

backed

0.97

 backing

0.81

interstitial

0.77

 backed

0.73

 militias

0.72

packs

0.72

ynamic

0.71

ragon

0.71

forcement

0.69

ollar

0.68

Activations Density 0.014%

phrases related to political or military backing or support

references to entities or groups supported by a governing body or organization

No Comments

No Known Activations