INDEX

Explanations

references to the initials "SS" at different levels of significance

oai_token-act-pair · gpt-3.5-turbo

references to specific medical or health-related concepts, particularly those associated with "SS"

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

staking

-0.89

tics

-0.87

Downloadha

-0.71

bidden

-0.71

tic

-0.69

 Advertisement

-0.68

 vain

-0.67

liness

-0.66

tical

-0.66

 stresses

-0.65

POSITIVE LOGITS

BM

0.99

avorite

0.89

emen

0.86

gt

0.84

agogue

0.83

MU

0.81

DP

0.80

SS

0.79

dinand

0.79

HT

0.78

Activations Density 0.013%

references to the initials "SS" at different levels of significance

references to specific medical or health-related concepts, particularly those associated with "SS"

No Comments

No Known Activations