INDEX

Explanations

phrases related to visual appearance or hypothetical scenarios

oai_token-act-pair · gpt-3.5-turbo

phrases discussing appearances or projections of how things will look in the future

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

cipl

-0.73

ukong

-0.72

ilts

-0.63

 QUEST

-0.61

lo

-0.61

rists

-0.59

ishable

-0.58

Jes

-0.58

Bac

-0.57

conn

-0.57

POSITIVE LOGITS

 nowadays

0.77

 today

0.75

 WITHOUT

0.73

 inside

0.72

 tonight

0.72

 AFTER

0.71

 BEFORE

0.71

 outside

0.69

 when

0.69

 before

0.69

Activations Density 0.056%

phrases related to visual appearance or hypothetical scenarios

phrases discussing appearances or projections of how things will look in the future

No Comments

No Known Activations