INDEX

Explanations

phrases where someone is being referred to or referenced

oai_token-act-pair · gpt-3.5-turbo

references or mentions of individuals or groups in a context

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

iem

-0.72

herer

-0.72

pite

-0.71

 depended

-0.66

 exploited

-0.66

oval

-0.64

ernel

-0.64

 relies

-0.63

works

-0.61

olitics

-0.61

POSITIVE LOGITS

gars

0.78

 phrases

0.76

ript

0.73

PsyNetMessage

0.70

çīĪ

0.70

 phrase

0.68

Ô

0.63

 è£ı

0.62

 deaf

0.62

 Topic

0.61

Activations Density 0.057%

phrases where someone is being referred to or referenced

references or mentions of individuals or groups in a context

No Comments

No Known Activations