INDEX

Explanations

the word "peninsula" in different contexts and locations

oai_token-act-pair · gpt-3.5-turbo

mentions of various geographic peninsulas

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

thumbnails

-0.81

nih

-0.70

 guts

-0.69

Interstitial

-0.69

ª

-0.69

¼

-0.69

ttle

-0.67

ī

-0.67

¯

-0.65

NL

-0.65

POSITIVE LOGITS

 peninsula

1.06

 Peninsula

1.04

insula

0.82

side

0.81

wide

0.73

conservancy

0.73

itect

0.73

 Tunisia

0.72

ampton

0.72

patrick

0.71

Activations Density 0.015%

the word "peninsula" in different contexts and locations

mentions of various geographic peninsulas

No Comments

No Known Activations