INDEX

Explanations

statements expressing opinions or suggestions

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate uncertainty or suggestion

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 Illusion

-0.65

unda

-0.61

 Elise

-0.60

 Sandwich

-0.60

Initialized

-0.60

 Harlem

-0.59

izabeth

-0.59

 Roll

-0.58

 Loud

-0.57

 Clear

-0.57

POSITIVE LOGITS

 someday

0.96

rosso

0.91

 forgiven

0.81

rued

0.75

omaly

0.75

ivably

0.74

iors

0.72

vier

0.70

 underestimated

0.70

TAIN

0.69

Activations Density 0.398%

statements expressing opinions or suggestions

phrases that indicate uncertainty or suggestion

No Comments

No Known Activations