INDEX

Explanations

positive or supportive statements, potentially in response to an action or situation

oai_token-act-pair · gpt-3.5-turbo

special characters and symbols punctuating textual expressions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 range

-0.80

 myster

-0.75

 invention

-0.75

 mathemat

-0.74

 mushroom

-0.73

 metic

-0.72

 usefulness

-0.71

 ranges

-0.71

 notor

-0.71

 habit

-0.71

POSITIVE LOGITS

ï¸ı

1.45

ONSORED

1.00

https

0.99

ttp

0.99

ðŁ

0.98

twitter

0.93

CNN

0.92

Â¯

0.92

Twitter

0.91

ðŁĺ

0.90

Activations Density 0.255%

positive or supportive statements, potentially in response to an action or situation

special characters and symbols punctuating textual expressions

No Comments

No Known Activations