INDEX

Explanations

words related to urging or encouraging someone to take action

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "prompt" and related variations in the context of instructions or requests

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 Nanto

-0.70

 Sect

-0.65

 Crescent

-0.64

 GEAR

-0.64

aples

-0.64

 Extrem

-0.63

=-=-=-=-=-=-=-=-

-0.63

achelor

-0.62

 Angels

-0.62

aredevil

-0.62

POSITIVE LOGITS

 prompt

1.41

 prompts

1.07

 Prompt

0.93

itude

0.92

 showc

0.89

tale

0.82

 warnings

0.77

 succession

0.76

 prompting

0.73

ously

0.73

Activations Density 0.005%

words related to urging or encouraging someone to take action

occurrences of the word "prompt" and related variations in the context of instructions or requests

No Comments

No Known Activations