INDEX

Explanations

phrases containing insults and accusations

oai_token-act-pair · gpt-3.5-turbo

derogatory comments or insults directed at individuals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

umerous

-0.76

ourse

-0.76

sequently

-0.74

However

-0.74

Additional

-0.73

etermined

-0.71

ufficient

-0.67

ilst

-0.67

orthy

-0.66

respective

-0.66

POSITIVE LOGITS

 fucking

1.29

 shit

1.25

 fuckin

1.22

 goddamn

1.20

 bullshit

1.17

 fucked

1.15

 crap

1.14

 dudes

1.12

 shitty

1.11

 fuck

1.09

Activations Density 1.414%

phrases containing insults and accusations

derogatory comments or insults directed at individuals

No Comments

No Known Activations