INDEX

Explanations

dates written in news article formats

oai_token-act-pair · gpt-3.5-turbo

dates and time references

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 tide

-0.81

luster

-0.70

 siph

-0.64

 appar

-0.64

 identifiable

-0.63

 unsuccessfully

-0.63

 correl

-0.62

 mater

-0.62

emonic

-0.61

 rejuven

-0.61

POSITIVE LOGITS

leases

0.89

eworld

0.87

 Retrieved

0.81

Associated

0.81

aughs

0.81

eport

0.79

 Accessed

0.76

Politics

0.75

humans

0.75

Fact

0.74

Activations Density 0.180%

dates written in news article formats

dates and time references

No Comments

No Known Activations