INDEX

Explanations

phrases related to winning awards, championships, titles, or games

oai_token-act-pair · gpt-3.5-turbo

mentions of wins and awards

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ngth

-0.83

onge

-0.73

 deviations

-0.68

ÃįÃį

-0.67

cised

-0.67

 ceilings

-0.67

isine

-0.67

igmat

-0.66

ofi

-0.64

idered

-0.64

POSITIVE LOGITS

 lottery

0.86

stakes

0.85

 Pulitzer

0.80

 prize

0.78

 outright

0.75

 prizes

0.72

 Sutherland

0.71

 landslide

0.70

 championship

0.70

 Sabha

0.69

Activations Density 0.155%

phrases related to winning awards, championships, titles, or games

mentions of wins and awards

No Comments

No Known Activations