INDEX

Explanations

mentions of locations and events related to political activities

oai_token-act-pair · gpt-3.5-turbo

mentions of political events and entities

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

âĢ

-0.81

by

-0.79

 withd

-0.73

â

-0.68

Ö¼

-0.66

By

-0.65

 âĸĪ

-0.61

ãĢ

-0.60

âĢł

-0.60

âĻ

-0.60

POSITIVE LOGITS

imeo

0.63

rompt

0.57

 complement

0.57

 contrasted

0.57

ĪĴ

0.55

cknow

0.54

 outweigh

0.54

 onwards

0.52

 playbook

0.52

entimes

0.52

Activations Density 0.903%

mentions of locations and events related to political activities

mentions of political events and entities

No Comments

No Known Activations