INDEX

Explanations

words related to something being valuable or important

oai_token-act-pair · gpt-3.5-turbo

mentions of things considered valuable

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

 Fever

-0.77

annis

-0.76

abi

-0.75

olog

-0.73

alis

-0.71

ulhu

-0.70

abel

-0.66

ologies

-0.66

central

-0.65

ilk

-0.65

POSITIVE LOGITS

 asset

0.97

 valuable

0.81

 trove

0.80

 contribution

0.79

 propositions

0.78

 proposition

0.77

 insights

0.77

Marginal

0.74

 assets

0.74

 possessions

0.73

Activations Density 0.021%

words related to something being valuable or important

mentions of things considered valuable

No Comments

No Known Activations