INDEX

Explanations

phrases related to investigations and dishonesty

oai_token-act-pair · gpt-3.5-turbo

instances of the character "ľ" in the text

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

raints

-0.76

matic

-0.75

 condem

-0.75

 Instr

-0.73

 ropes

-0.72

rouse

-0.70

ulators

-0.70

 grips

-0.69

 organisers

-0.68

 tyres

-0.68

POSITIVE LOGITS

ï¸ı

1.18

âĶĢâĶĢ

1.17

âĶĢâĶĢâĶĢâĶĢâĶĢâĶĢâĶĢâĶĢ

0.93

conom

0.90

×Ķ

0.85

Ł

0.85

ĸ

0.83

Ĩ

0.83

ł

0.83

¸

0.82

Activations Density 0.158%

phrases related to investigations and dishonesty

instances of the character "ľ" in the text

No Comments

No Known Activations