INDEX

Explanations

phrases indicating a command or instruction

oai_token-act-pair · gpt-3.5-turbo

negative imperatives or prohibition phrases

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 ancest

-0.72

CVE

-0.63

 Reloaded

-0.62

 Frie

-0.61

milo

-0.56

 behavi

-0.56

parser

-0.56

gnu

-0.55

wrapper

-0.55

 spirited

-0.54

POSITIVE LOGITS

 hesitate

0.90

 forget

0.89

 bother

0.87

 expect

0.86

Í

0.85

intend

0.81

 necessarily

0.80

 CARE

0.80

ude

0.80

erest

0.77

Activations Density 0.057%

phrases indicating a command or instruction

negative imperatives or prohibition phrases

No Comments

No Known Activations