INDEX

Explanations

questions or requests directed towards someone

oai_token-act-pair · gpt-3.5-turbo

instances of the word "asked."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

Ĥ¬

-0.83

 Scouting

-0.74

marine

-0.73

âĸ¬

-0.67

âĶĢâĶĢ

-0.65

cyclop

-0.60

\\\\

-0.60

smoking

-0.60

mit

-0.60

ordinate

-0.60

POSITIVE LOGITS

 rhet

1.18

 questions

1.03

 probing

0.98

 politely

0.89

him

0.87

 forgiveness

0.85

 asked

0.85

naires

0.84

ioned

0.82

 permission

0.81

Activations Density 0.044%

questions or requests directed towards someone

instances of the word "asked."

No Comments

No Known Activations