INDEX

Explanations

mentions of geographical locations and associated words

oai_token-act-pair · gpt-3.5-turbo

references to social issues and institutions related to politics and education

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

escription

-0.63

anwhile

-0.56

 tweeting

-0.56

 Rhod

-0.54

Tai

-0.53

 Notting

-0.51

uay

-0.51

CLOSE

-0.51

 condem

-0.51

!:

-0.51

POSITIVE LOGITS

 differs

1.24

 varies

1.17

 exceeds

1.17

 depends

1.17

 stems

1.12

 depended

1.09

 hinges

1.08

 outweigh

1.05

 correlates

1.04

 resembles

1.03

Activations Density 0.313%

mentions of geographical locations and associated words

references to social issues and institutions related to politics and education

No Comments

No Known Activations