INDEX

Explanations

words related to legal or authoritative rules or examples

oai_token-act-pair · gpt-3.5-turbo

instances of the word "precedent" and its variations

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

wagen

-0.73

ococ

-0.70

otrop

-0.70

ocations

-0.69

yne

-0.69

porting

-0.68

yan

-0.68

rug

-0.68

ocated

-0.67

kie

-0.67

POSITIVE LOGITS

 precedent

1.16

 preced

0.92

 precedence

0.88

DragonMagazine

0.88

chall

0.74

©¶æ¥µ

0.72

ĸļ

0.72

peer

0.69

insula

0.65

Rule

0.64

Activations Density 0.013%

words related to legal or authoritative rules or examples

instances of the word "precedent" and its variations

No Comments

No Known Activations