INDEX

Explanations

terms related to legal documents or procedures

oai_token-act-pair · gpt-3.5-turbo

instances of the word "writ" and related terms like "operator."

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

theless

-0.82

 Fortress

-0.77

arya

-0.70

»Ĵ

-0.69

riad

-0.69

tones

-0.66

 Townsend

-0.64

beard

-0.64

istant

-0.63

illon

-0.63

POSITIVE LOGITS

eworks

0.95

 writ

0.90

ework

0.88

acion

0.88

atical

0.77

eno

0.75

natureconservancy

0.74

 ARTICLE

0.70

 membr

0.69

emic

0.67

Activations Density 0.035%

terms related to legal documents or procedures

instances of the word "writ" and related terms like "operator."

No Comments

No Known Activations