INDEX

Explanations

words related to professions, such as attorney, lawyer, investigator, and author

oai_token-act-pair · gpt-3.5-turbo

references to legal professionals and family relationships

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

twitch

-0.67

"]=>

-0.64

rising

-0.64

attering

-0.64

avers

-0.64

verage

-0.62

uchin

-0.62

flush

-0.62

discrimination

-0.62

lems

-0.62

POSITIVE LOGITS

 Brendan

0.99

 Andrew

0.99

 Jamie

0.98

 Nathaniel

0.97

 Jonah

0.97

Ian

0.97

 Roger

0.96

 Linda

0.95

 Patrick

0.95

 Geoffrey

0.95

Activations Density 0.202%

words related to professions, such as attorney, lawyer, investigator, and author

references to legal professionals and family relationships

No Comments

No Known Activations