INDEX

Explanations

words related to components or parts

oai_token-act-pair · gpt-3.5-turbo

terms related to components within a technical or product context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

kers

-0.81

chie

-0.76

ker

-0.73

standing

-0.72

STER

-0.70

geon

-0.69

stretched

-0.68

Calling

-0.67

ield

-0.66

à¦

-0.66

POSITIVE LOGITS

 components

1.10

 component

0.96

 parts

0.93

 Components

0.92

 subsystem

0.86

onents

0.84

 Component

0.78

 additive

0.76

ancial

0.75

 elements

0.74

Activations Density 0.011%

words related to components or parts

terms related to components within a technical or product context

No Comments

No Known Activations