INDEX

Explanations

words related to legal consequences and judgments

oai_token-act-pair · gpt-3.5-turbo

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

TC

-0.73

 Classic

-0.70

FA

-0.70

 Pros

-0.69

meta

-0.67

 Plays

-0.66

 Ducks

-0.66

tr

-0.66

TS

-0.66

EF

-0.65

POSITIVE LOGITS

 imprisonment

3.51

 incarceration

3.02

 imprison

2.79

 imprisoned

2.51

 incarcer

2.38

 incarcerated

2.18

prison

1.94

arcer

1.93

 prison

1.92

 confinement

1.74

Activations Density 0.021%

words related to legal consequences and judgments

No Comments

No Known Activations

words related to legal consequences and judgments

No Comments

No Known Activations