INDEX

Explanations

objects or items that are piled, stacked, or arranged in layers

oai_token-act-pair · gpt-3.5-turbo

references to stacking or piling items

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

nton

-0.87

tained

-0.72

shire

-0.72

gl

-0.71

kiss

-0.68

âĵĺ

-0.67

hood

-0.67

estamp

-0.66

bies

-0.66

POSITIVE LOGITS

 stacked

0.99

 piles

0.91

 shelves

0.87

 stacks

0.78

 stacking

0.78

river

0.77

 racks

0.75

 unbeliev

0.74

 boxes

0.74

 piled

0.72

Activations Density 0.020%

objects or items that are piled, stacked, or arranged in layers

references to stacking or piling items

No Comments

No Known Activations