INDEX

Explanations

phrases related to government actions or statements

oai_token-act-pair · gpt-3.5-turbo

symbols or characters resembling the "â" character

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

zag

-0.67

 maxim

-0.67

 subur

-0.67

 scatter

-0.67

 radar

-0.66

 minim

-0.64

 scene

-0.64

 obser

-0.63

 spinning

-0.63

ctors

-0.62

POSITIVE LOGITS

£

0.95

_>

0.88

Â¢

0.82

âĢł

0.80

Â¯

0.80

there

0.78

sure

0.77

º

0.75

âĢ

0.75

>[

0.75

Activations Density 0.375%

phrases related to government actions or statements

symbols or characters resembling the "â" character

No Comments

No Known Activations