INDEX

Explanations

locations or objects related to physical structures

oai_token-act-pair · gpt-3.5-turbo

locations and settings associated with incidents or events

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

 sexes

-0.81

Ps

-0.75

 positives

-0.74

phas

-0.69

 Parties

-0.66

 prizes

-0.66

ernels

-0.63

 feats

-0.63

 tokens

-0.63

 pics

-0.63

POSITIVE LOGITS

 courtyard

0.83

robe

0.81

 riddled

0.79

 balcony

0.76

 overlooking

0.76

 blender

0.75

 alley

0.74

 cafÃ©

0.73

 hallway

0.73

 airst

0.72

Activations Density 0.373%

locations or objects related to physical structures

locations and settings associated with incidents or events

No Comments

No Known Activations