INDEX

Explanations

editor's notes within articles

oai_token-act-pair · gpt-3.5-turbo

possessive forms indicating ownership or editorial attributions

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

ĪĴ

-0.79

ansas

-0.75

vironment

-0.72

facts

-0.68

wikipedia

-0.67

imedia

-0.67

esm

-0.66

USD

-0.66

flix

-0.66

quished

-0.64

POSITIVE LOGITS

own

0.83

 inability

0.80

 remorse

0.77

 grasp

0.70

 insistence

0.70

 Guild

0.70

 ability

0.69

 penchant

0.68

 daughter

0.68

 Wife

0.68

Activations Density 0.094%

editor's notes within articles

possessive forms indicating ownership or editorial attributions

No Comments

No Known Activations