INDEX

Explanations

terms related to demonization or negative portrayal of certain groups or individuals

oai_token-act-pair · gpt-3.5-turbo

instances of the word "demon" in various 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

ippi

-0.72

RAFT

-0.68

Ã¥

-0.66

 proble

-0.66

 Seym

-0.65

jri

-0.65

ILLE

-0.63

ills

-0.63

hiba

-0.62

 Crosby

-0.62

POSITIVE LOGITS

stration

1.20

iac

1.08

ising

0.95

oid

0.93

ises

0.92

ormal

0.92

izing

0.88

ised

0.88

ization

0.86

izes

0.84

Activations Density 0.015%

terms related to demonization or negative portrayal of certain groups or individuals

instances of the word "demon" in various contexts

No Comments

No Known Activations