INDEX

Explanations

references to tiles or tiling in construction or building contexts

oai_token-act-pair · gpt-3.5-turbo

references to tiles and their quantity or condition

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ulner

-0.92

ntil

-0.73

ebin

-0.72

 largeDownload

-0.72

avez

-0.72

jri

-0.72

ournal

-0.71

etheus

-0.71

ivia

-0.70

hovah

-0.68

POSITIVE LOGITS

 tiles

1.10

 Tile

0.96

 tile

0.96

 sheets

0.86

 grids

0.83

sheet

0.79

 mosa

0.79

 decoration

0.79

 sheet

0.76

 Walls

0.76

Activations Density 0.014%

references to tiles or tiling in construction or building contexts

references to tiles and their quantity or condition

No Comments

No Known Activations