INDEX

Explanations

actions involving processing, handling, or transforming materials

oai_token-act-pair · gpt-3.5-turbo

verbs indicating processes or sequences of actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 inaction

-0.73

 Failure

-0.72

 Matters

-0.71

Atk

-0.70

 Lack

-0.70

 Dare

-0.69

comings

-0.68

isSpecialOrderable

-0.68

 detrim

-0.67

 Immunity

-0.66

POSITIVE LOGITS

 transported

1.23

 converted

1.22

 shipped

1.19

 scanned

1.18

 transformed

1.18

 placed

1.18

 rotated

1.17

 inspected

1.17

 sprayed

1.17

 subjected

1.17

Activations Density 0.173%

actions involving processing, handling, or transforming materials

verbs indicating processes or sequences of actions

No Comments

No Known Activations