INDEX

Explanations

technical terms related to a specific brand or company

oai_token-act-pair · gpt-3.5-turbo

references to a specific brand or model of product and associated details

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

kaya

-0.93

selves

-0.82

Magikarp

-0.79

self

-0.76

iyah

-0.73

achu

-0.71

edly

-0.70

enko

-0.69

negie

-0.69

bian

-0.69

POSITIVE LOGITS

OC

0.84

 bean

0.72

ovo

0.71

SC

0.71

 Bean

0.71

arios

0.69

sequence

0.68

ormal

0.67

FP

0.65

SE

0.65

Activations Density 0.013%

technical terms related to a specific brand or company

references to a specific brand or model of product and associated details

No Comments

No Known Activations