INDEX

Explanations

mentions of events or actions that started or initiated something new

oai_token-act-pair · gpt-3.5-turbo

instances of the word "started" in various contexts

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

cit

-0.81

etry

-0.75

acho

-0.73

ighth

-0.73

ethy

-0.73

âĨĳ

-0.72

alted

-0.71

ugs

-0.69

ses

-0.69

ingly

-0.69

POSITIVE LOGITS

 anew

0.93

 raining

0.77

 PRESS

0.76

OCK

0.75

ŃĶ

0.74

 airing

0.73

 fuss

0.71

 circulating

0.71

 behaving

0.71

 dating

0.67

Activations Density 0.056%

mentions of events or actions that started or initiated something new

instances of the word "started" in various contexts

No Comments

No Known Activations