INDEX

Explanations

mentions of specific numerical levels

oai_token-act-pair · gpt-3.5-turbo

references to "levels" related to various substances or conditions

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

ILLE

-0.76

Untitled

-0.74

ãĥĨãĤ£

-0.71

nell

-0.69

sov

-0.68

rontal

-0.68

 Andromeda

-0.67

INA

-0.67

ãĥĹ

-0.67

enegger

-0.65

POSITIVE LOGITS

 levels

0.89

 detectable

0.87

 concentrations

0.83

baugh

0.83

 amounts

0.79

 quantities

0.74

afety

0.74

cius

0.73

 attained

0.72

 fluctuations

0.70

Activations Density 0.026%

mentions of specific numerical levels

references to "levels" related to various substances or conditions

No Comments

No Known Activations