INDEX

Explanations

positive adverbs that describe emotions or actions

oai_token-act-pair · gpt-3.5-turbo

adverbs that describe manner or attitude

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

 strictly

-0.75

YL

-0.74

senal

-0.72

ãĥķãĤ©

-0.70

 truly

-0.68

LER

-0.67

 REALLY

-0.67

 clearly

-0.67

 really

-0.67

HS

-0.66

POSITIVE LOGITS

 awaited

1.11

 awaits

0.94

 accepted

0.93

 embraced

0.92

 complied

0.91

 awaiting

0.91

 await

0.88

 conceded

0.86

 dispatched

0.86

 acquies

0.85

Activations Density 0.101%

positive adverbs that describe emotions or actions

adverbs that describe manner or attitude

No Comments

No Known Activations