INDEX

Explanations

proper nouns related to various individuals and locations

oai_token-act-pair · gpt-3.5-turbo

proper nouns and names

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 telecommunications

-0.56

 appealing

-0.55

 lucrative

-0.54

 midway

-0.53

 alongside

-0.52

 safegu

-0.52

 proactive

-0.52

 prominent

-0.52

 ostensibly

-0.52

 nurturing

-0.51

POSITIVE LOGITS

Walker

0.73

Knight

0.72

Wiki

0.69

Points

0.69

Clock

0.68

Squ

0.68

:=

0.67

Girl

0.67

Style

0.67

Bomb

0.66

Activations Density 0.500%

proper nouns related to various individuals and locations

proper nouns and names

No Comments

No Known Activations