INDEX

Explanations

references to specific abbreviations or acronyms beginning with "TR"

oai_token-act-pair · gpt-3.5-turbo

references to the abbreviation "TR" in various contexts

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

 actionGroup

-0.80

manship

-0.77

eers

-0.74

esville

-0.71

ãģ®éŃĶ

-0.71

lace

-0.69

creen

-0.68

âĸ¬âĸ¬

-0.68

comes

-0.67

soever

-0.67

POSITIVE LOGITS

ACK

1.05

UTH

0.97

ainer

0.94

ACT

0.92

acing

0.91

ACE

0.89

IP

0.89

umble

0.88

usted

0.85

ADE

0.83

Activations Density 0.006%

references to specific abbreviations or acronyms beginning with "TR"

references to the abbreviation "TR" in various contexts

No Comments

No Known Activations