INDEX

Explanations

mentions of physical locations or buildings

oai_token-act-pair · gpt-3.5-turbo

references to different types of complexes or compounds

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

 paralysis

-0.69

ãģĦ

-0.67

Force

-0.66

lime

-0.65

confidence

-0.63

yon

-0.61

reme

-0.61

alys

-0.60

ohan

-0.59

 Expression

-0.59

POSITIVE LOGITS

pole

0.94

 overlooking

0.90

tyard

0.84

 situated

0.82

 premises

0.81

 occupancy

0.81

 located

0.80

querque

0.79

 Hotel

0.77

 Afric

0.77

Activations Density 0.196%

mentions of physical locations or buildings

references to different types of complexes or compounds

No Comments

No Known Activations