INDEX

Explanations

instances where actions are being attributed to individuals or entities

oai_token-act-pair · gpt-3.5-turbo

instances of the verb "puts."

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

NI

-0.67

omorphic

-0.65

oooooooooooooooo

-0.59

gyn

-0.59

ISSION

-0.57

adobe

-0.56

iform

-0.55

 Alleg

-0.54

 surviving

-0.53

 commercially

-0.52

POSITIVE LOGITS

 puts

3.11

put

1.52

 pushes

1.50

 shuts

1.45

 gives

1.44

 throws

1.42

 sends

1.41

 brings

1.39

 lays

1.38

 translates

1.38

Activations Density 0.012%

instances where actions are being attributed to individuals or entities

instances of the verb "puts."

No Comments

No Known Activations