INDEX

Explanations

legal terms or actions related to law enforcement and legality

oai_token-act-pair · gpt-3.5-turbo

concepts related to legality, specifically focusing on the terms "lawful" and "unlawful."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Tycoon

-0.84

bol

-0.82

olog

-0.75

ocl

-0.74

 Solitaire

-0.73

culosis

-0.71

 Atlas

-0.69

lasses

-0.68

pocket

-0.68

 Anxiety

-0.67

POSITIVE LOGITS

theless

0.96

rontal

0.84

00200000

0.82

cy

0.81

 unlawful

0.74

 unfocusedRange

0.74

 interference

0.71

 Liberties

0.70

cipled

0.69

 possession

0.69

Activations Density 0.020%

legal terms or actions related to law enforcement and legality

concepts related to legality, specifically focusing on the terms "lawful" and "unlawful."

No Comments

No Known Activations