INDEX

Explanations

references to specific education or training programs

oai_token-act-pair · gpt-3.5-turbo

references to educational or instructional categories

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Leaks

-0.80

agate

-0.75

erry

-0.70

lez

-0.66

dor

-0.64

redo

-0.64

 Flip

-0.62

oil

-0.61

pire

-0.61

jer

-0.61

POSITIVE LOGITS

 classes

3.96

 Classes

3.33

classes

2.93

 class

2.59

Class

2.10

class

1.98

 Class

1.96

 courses

1.78

 CLASS

1.75

 subclass

1.62

Activations Density 0.015%

references to specific education or training programs

references to educational or instructional categories

No Comments

No Known Activations