INDEX

Explanations

phrases related to legal matters, specifically parole

oai_token-act-pair · gpt-3.5-turbo

terms related to parole and its implications in the criminal justice system

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

sonian

-0.95

advertisement

-0.74

iance

-0.69

 Grimm

-0.69

omen

-0.65

ractive

-0.65

 MÃ¼

-0.64

ies

-0.63

DW

-0.62

Jewish

-0.61

POSITIVE LOGITS

 parole

1.25

reon

0.94

 probation

0.91

 confir

0.86

 Ribbon

0.80

ctors

0.80

 clearance

0.79

 eligibility

0.79

 versa

0.78

 revocation

0.78

Activations Density 0.006%

phrases related to legal matters, specifically parole

terms related to parole and its implications in the criminal justice system

No Comments

No Known Activations