INDEX

Explanations

words related to the concept of locking or being locked

oai_token-act-pair · gpt-3.5-turbo

references to being confined or restricted

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 enrich

-0.69

Nu

-0.65

 outing

-0.64

Pep

-0.64

 Express

-0.64

 Cout

-0.63

 inspir

-0.62

 contributions

-0.62

iasm

-0.62

anson

-0.62

POSITIVE LOGITS

locked

4.43

locks

2.39

locking

2.10

lock

1.86

LOCK

1.54

 Locked

1.21

 locked

1.21

aligned

1.11

 unlocked

1.10

Lock

1.08

Activations Density 0.017%

words related to the concept of locking or being locked

references to being confined or restricted

No Comments

No Known Activations