INDEX

Explanations

references to different levels of measurement or quantity

oai_token-act-pair · gpt-3.5-turbo

references to varying levels of measurement or quantities

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

nee

-0.77

 towed

-0.75

cele

-0.75

widget

-0.74

closed

-0.72

jen

-0.71

oubt

-0.65

ridge

-0.64

ida

-0.63

 Warrant

-0.63

POSITIVE LOGITS

 levels

3.81

levels

2.87

 Levels

2.85

 level

2.41

 LEVEL

2.16

 concentrations

1.91

 Level

1.73

Level

1.72

level

1.69

 heights

1.67

Activations Density 0.017%

references to different levels of measurement or quantity

references to varying levels of measurement or quantities

No Comments

No Known Activations