INDEX

Explanations

phrases related to conflicts and negotiations between two opposing sides

oai_token-act-pair · gpt-3.5-turbo

references to opposing parties or groups within a conflict

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

nce

-0.67

 Miko

-0.63

Blade

-0.61

Delivery

-0.61

glomer

-0.60

 Rapids

-0.58

 renovations

-0.57

Ost

-0.57

 Mansion

-0.57

entric

-0.56

POSITIVE LOGITS

 alike

1.04

 vying

0.93

 sides

0.93

 parties

0.90

 concede

0.88

cale

0.88

 thereof

0.88

 Parties

0.86

 equally

0.85

 agree

0.82

Activations Density 0.084%

phrases related to conflicts and negotiations between two opposing sides

references to opposing parties or groups within a conflict

No Comments

No Known Activations