INDEX

Explanations

postal codes or other numerical identifiers used in different locations

oai_token-act-pair · gpt-3.5-turbo

instances of significant numerical references or categorizations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Spectre

-0.72

 veter

-0.69

 Frenzy

-0.69

 Versus

-0.68

lyak

-0.67

 Contra

-0.67

 perks

-0.62

 Flesh

-0.61

 perk

-0.61

 disadvant

-0.61

POSITIVE LOGITS

English

0.99

isine

0.95

abulary

0.86

"}],"

0.81

Attach

0.79

Posted

0.79

Population

0.78

podcast

0.75

Subject

0.75

dayName

0.74

Activations Density 0.082%

postal codes or other numerical identifiers used in different locations

instances of significant numerical references or categorizations

No Comments

No Known Activations