INDEX

Explanations

references to a physical chain, like in supply chains or a sequence of connected objects

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "chain" in various contexts

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

MAT

-0.78

 Kepler

-0.70

 Lowell

-0.70

 Lavrov

-0.69

 Loans

-0.68

sburgh

-0.68

Hue

-0.67

Dob

-0.66

Marginal

-0.65

bernatorial

-0.65

POSITIVE LOGITS

 chain

1.29

chain

1.19

Chain

1.12

 chains

1.03

chains

0.88

rings

0.85

alysis

0.83

worm

0.83

 Chain

0.83

wal

0.82

Activations Density 0.008%

references to a physical chain, like in supply chains or a sequence of connected objects

occurrences of the word "chain" in various contexts

No Comments

No Known Activations