INDEX

Explanations

features related to the durability or playability of objects or products

oai_token-act-pair · gpt-3.5-turbo

terms related to product features and characteristics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 zeal

-0.75

invest

-0.69

 neglect

-0.68

FT

-0.67

hoe

-0.66

hai

-0.66

mist

-0.65

 Overt

-0.65

cyn

-0.65

 Studies

-0.64

POSITIVE LOGITS

 usable

2.32

 removable

2.24

 customizable

2.09

 durable

2.08

 playable

2.06

 adjustable

2.04

 reversible

1.94

 readable

1.93

 reusable

1.90

 wearable

1.80

Activations Density 0.049%

features related to the durability or playability of objects or products

terms related to product features and characteristics

No Comments

No Known Activations