INDEX

Explanations

superlatives related to levels or ranks

oai_token-act-pair · gpt-3.5-turbo

phrases indicating the highest levels or rankings

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

fy

-0.76

abel

-0.66

arry

-0.65

lins

-0.64

bug

-0.63

lessly

-0.62

zilla

-0.61

kin

-0.59

wen

-0.59

Doc

-0.57

POSITIVE LOGITS

 highest

3.58

highest

2.84

 lowest

2.68

 Highest

2.34

 widest

2.05

 heaviest

1.87

 greatest

1.85

 tallest

1.81

 strongest

1.79

Hig

1.78

Activations Density 0.014%

superlatives related to levels or ranks

phrases indicating the highest levels or rankings

No Comments

No Known Activations