INDEX

Explanations

phrases related to news headlines or current events

oai_token-act-pair · gpt-3.5-turbo

mentions of locations and political 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

anwhile

-0.67

itored

-0.54

 respectively

-0.52

ashington

-0.51

 local

-0.50

 Local

-0.48

srf

-0.47

igible

-0.47

redited

-0.47

named

-0.46

POSITIVE LOGITS

coin

0.56

guy

0.56

sag

0.53

 sucks

0.50

 guys

0.50

 devs

0.49

Coin

0.47

!)

0.46

gamer

0.45

istor

0.45

Activations Density 2.571%

phrases related to news headlines or current events

mentions of locations and political contexts

No Comments

No Known Activations