INDEX

Explanations

phrases introducing additional information

oai_token-act-pair · gpt-3.5-turbo

references to documents or reports that characterize or provide information about a particular subject or event

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

mma

-0.62

 guessing

-0.62

Â·

-0.61

 ignorance

-0.61

Pot

-0.61

 Paradox

-0.59

icidal

-0.59

 believing

-0.59

ocent

-0.59

Disable

-0.58

POSITIVE LOGITS

 launched

0.93

 fielded

0.90

released

0.89

 overseen

0.86

 jointly

0.86

leased

0.86

 convened

0.84

 released

0.83

 published

0.82

 spearheaded

0.82

Activations Density 0.349%

phrases introducing additional information

references to documents or reports that characterize or provide information about a particular subject or event

No Comments

No Known Activations