INDEX

Explanations

phrases relating to spatial directions, particularly downward movement

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate movement or positioning along a path or road

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

akeru

-0.89

orth

-0.84

ilver

-0.82

ItemTracker

-0.79

riks

-0.77

construct

-0.75

wu

-0.75

asus

-0.74

itaire

-0.74

aram

-0.73

POSITIVE LOGITS

 throats

0.88

 weeds

0.82

 stairs

0.81

 hallway

0.76

dden

0.76

 road

0.74

 dumps

0.72

 slopes

0.72

 rabbit

0.71

 freeway

0.69

Activations Density 0.125%

phrases relating to spatial directions, particularly downward movement

phrases that indicate movement or positioning along a path or road

No Comments

No Known Activations