INDEX

Explanations

words related to legal situations and actions

oai_token-act-pair · gpt-3.5-turbo

legal terms and concepts related to crime and justice

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

âĢ

-0.84

ãĥ»

-0.76

*)

-0.71

 âĶĤ

-0.70

*,

-0.69

 âĢ

-0.68

à

-0.67

}{

-0.66

âĢł

-0.65

â

-0.65

POSITIVE LOGITS

to

0.75

 thereto

0.53

to

0.52

 condemning

0.51

 redes

0.50

To

0.50

 justifying

0.50

heit

0.49

 denying

0.49

 herein

0.48

Activations Density 0.928%

words related to legal situations and actions

legal terms and concepts related to crime and justice

No Comments

No Known Activations