INDEX

Explanations

words related to acts of violence or killing

oai_token-act-pair · gpt-3.5-turbo

references to animal and human slaughter

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 Princ

-0.65

DCS

-0.64

ioxide

-0.63

 Crom

-0.63

 Ronaldo

-0.61

 charism

-0.59

 Eucl

-0.59

annis

-0.59

aucas

-0.58

 Richards

-0.57

POSITIVE LOGITS

 Slaughter

1.19

houses

1.16

 slaughter

1.00

house

0.97

ãĥ¼ãĤ¯

0.94

\\\\\\\\

0.93

quished

0.91

ificial

0.89

 slaughtered

0.84

eful

0.81

Activations Density 0.012%

words related to acts of violence or killing

references to animal and human slaughter

No Comments

No Known Activations