INDEX

Explanations

phrases indicating a level of concern or consideration

oai_token-act-pair · gpt-3.5-turbo

expressions indicating concern or desire

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

pload

-0.67

omas

-0.59

ware

-0.58

Ada

-0.56

 standoff

-0.56

 Cance

-0.56

 worry

-0.55

 Heavy

-0.54

 Rusty

-0.54

 âĹı

-0.54

POSITIVE LOGITS

alties

0.73

liest

0.70

FontSize

0.69

 practicable

0.69

lege

0.67

imum

0.65

isd

0.65

eeds

0.64

soType

0.61

 weakest

0.61

Activations Density 0.379%

phrases indicating a level of concern or consideration

expressions indicating concern or desire

No Comments

No Known Activations