INDEX

Explanations

phrases related to unspecified or to-be-determined information

oai_token-act-pair · gpt-3.5-turbo

terms related to unknown entities or statuses

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

grain

-0.83

tons

-0.81

romy

-0.77

Cooldown

-0.77

APH

-0.77

mean

-0.77

floor

-0.77

anon

-0.76

iven

-0.76

nas

-0.74

POSITIVE LOGITS

ecided

0.92

theless

0.88

 Turtle

0.79

TBD

0.76

 Guam

0.74

ulhu

0.72

 Hurricanes

0.71

 Valkyrie

0.68

TION

0.68

 Hedge

0.66

Activations Density 0.016%

phrases related to unspecified or to-be-determined information

terms related to unknown entities or statuses

No Comments

No Known Activations