INDEX

Explanations

dates or time-related expressions

oai_token-act-pair · gpt-3.5-turbo

references to timeline indicators or dates

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

estern

-0.85

alach

-0.78

aiman

-0.77

pmwiki

-0.73

obal

-0.72

 cytok

-0.71

SpaceEngineers

-0.70

iami

-0.70

ammy

-0.69

assad

-0.68

POSITIVE LOGITS

abouts

0.71

date

0.70

JD

0.69

.(

0.65

 Feast

0.64

APH

0.63

geist

0.63

dule

0.63

 date

0.63

Versions

0.63

Activations Density 0.013%

dates or time-related expressions

references to timeline indicators or dates

No Comments

No Known Activations