INDEX

Explanations

references to physical attributes or locations

oai_token-act-pair · gpt-3.5-turbo

references to physical presence or characteristics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

Gos

-0.67

iners

-0.66

 Spot

-0.66

Pont

-0.65

 Gould

-0.65

ills

-0.64

 Bank

-0.64

ered

-0.63

edin

-0.63

 Maxim

-0.63

POSITIVE LOGITS

physical

1.02

anguage

0.96

 manifestation

0.92

 cryst

0.90

 physical

0.88

 cumbers

0.87

 altercation

0.86

 manifestations

0.82

istic

0.82

ity

0.81

Activations Density 0.011%

references to physical attributes or locations

references to physical presence or characteristics

No Comments

No Known Activations