INDEX

Explanations

phrases related to harmful acts or actions

oai_token-act-pair · gpt-3.5-turbo

words related to acts of violence and their consequences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

arest

-0.72

 Plat

-0.70

aver

-0.67

oult

-0.67

pole

-0.66

iHUD

-0.66

hack

-0.65

 therapy

-0.64

oret

-0.63

arro

-0.61

POSITIVE LOGITS

 perpetrated

1.07

 committing

0.92

 withd

0.91

interstitial

0.90

 committed

0.86

ahime

0.84

 impunity

0.82

ãĥ¼ãĥĨ

0.80

 heinous

0.80

0.78

Activations Density 0.007%

phrases related to harmful acts or actions

words related to acts of violence and their consequences

No Comments

No Known Activations