INDEX

Explanations

mentions of physical space or spatial concepts

oai_token-act-pair · gpt-3.5-turbo

references to physical space or capacity

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

usted

-0.77

iamond

-0.77

ctive

-0.74

otine

-0.69

ku

-0.68

cause

-0.68

tl

-0.67

venge

-0.66

kai

-0.66

 decap

-0.64

POSITIVE LOGITS

 space

0.93

 spaces

0.90

 shuttle

0.81

 constraints

0.79

 habitats

0.76

bars

0.75

 occupied

0.74

kat

0.73

 Layout

0.73

ngth

0.72

Activations Density 0.028%

mentions of physical space or spatial concepts

references to physical space or capacity

No Comments

No Known Activations