INDEX

Explanations

words related to hostile or violent behaviors

oai_token-act-pair · gpt-3.5-turbo

references to aggression

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

CLE

-0.71

Lay

-0.71

 Plain

-0.70

SIM

-0.69

cloth

-0.68

Catalog

-0.67

ctors

-0.65

ird

-0.65

plain

-0.65

atch

-0.64

POSITIVE LOGITS

 aggression

3.54

 aggress

2.46

 hostility

1.69

 provocation

1.65

 aggressive

1.61

aggressive

1.56

Agg

1.44

 violence

1.43

 arrogance

1.38

 intimidation

1.38

Activations Density 0.027%

words related to hostile or violent behaviors

references to aggression

No Comments

No Known Activations