INDEX

Explanations

descriptions of political figures and their affiliations

oai_token-act-pair · gpt-3.5-turbo

titles and roles related to political figures and their affiliations

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 VIDEOS

-0.78

 ACTIONS

-0.66

Â¯Â¯

-0.63

ibraries

-0.62

 UNIVERS

-0.57

 Locations

-0.56

 Adults

-0.56

PLA

-0.55

LESS

-0.55

 NEWS

-0.54

POSITIVE LOGITS

 whom

1.35

who

1.29

 whose

1.25

whose

1.09

who

1.08

 nicknamed

1.00

 credited

0.97

 tasked

0.96

 accused

0.96

 responsible

0.95

Activations Density 0.165%

descriptions of political figures and their affiliations

titles and roles related to political figures and their affiliations

No Comments

No Known Activations