INDEX

Explanations

legal and judicial terminology

oai_token-act-pair · gpt-3.5-turbo

instances of specific characters or symbols

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

geries

-0.65

 unconscious

-0.64

è¦ļéĨĴ

-0.62

 abstinence

-0.62

 ducks

-0.61

 bluff

-0.61

 scissors

-0.61

 undecided

-0.59

 assemb

-0.59

berman

-0.59

POSITIVE LOGITS

tml

0.77

 Ensure

0.74

tab

0.73

vor

0.73

externalActionCode

0.73

Compare

0.73

âĶĢ

0.72

 =================================================================

0.71

 Likewise

0.70

İ

0.69

Activations Density 0.259%

legal and judicial terminology

instances of specific characters or symbols

No Comments

No Known Activations