INDEX

Explanations

phrases indicating extremes or exaggeration

oai_token-act-pair · gpt-3.5-turbo

expressions that emphasize extent or intensity

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

earing

-0.68

ufficient

-0.67

ractive

-0.64

pection

-0.64

iest

-0.64

unct

-0.63

ashington

-0.62

idable

-0.62

lement

-0.61

olen

-0.60

POSITIVE LOGITS

 overboard

1.19

 furthe

0.94

 downhill

0.81

 smoothly

0.80

WARD

0.78

ast

0.77

 farther

0.76

 unnoticed

0.76

EStreamFrame

0.71

nova

0.71

Activations Density 0.079%

phrases indicating extremes or exaggeration

expressions that emphasize extent or intensity

No Comments

No Known Activations