INDEX

Explanations

references to hierarchical positions or classifications

oai_token-act-pair · gpt-3.5-turbo

references to hierarchical positions or ranks within organizations

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

argo

-0.72

reens

-0.72

ogle

-0.68

credit

-0.68

agic

-0.67

pmwiki

-0.66

orescence

-0.66

reb

-0.65

Pwr

-0.65

berra

-0.63

POSITIVE LOGITS

 insign

1.04

nance

0.84

loader

0.80

quo

0.80

ings

0.78

ific

0.77

enance

0.76

çĦ

0.75

ename

0.74

leading

0.74

Activations Density 0.042%

references to hierarchical positions or classifications

references to hierarchical positions or ranks within organizations

No Comments

No Known Activations