INDEX

Explanations

words related to physical environment or location

oai_token-act-pair · gpt-3.5-turbo

mentions of the concept of "landscape" in various contexts

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

ibo

-0.83

etts

-0.81

ulate

-0.78

uster

-0.77

uana

-0.77

efficients

-0.77

gements

-0.76

abad

-0.76

unker

-0.76

ule

-0.76

POSITIVE LOGITS

scape

0.94

 Rover

0.90

ttes

0.86

 landscape

0.84

 landscapes

0.78

XT

0.74

 wallpaper

0.73

 Photographer

0.71

 acre

0.71

 backdrop

0.70

Activations Density 0.008%

words related to physical environment or location

mentions of the concept of "landscape" in various contexts

No Comments

No Known Activations