INDEX

Explanations

superlatives related to effectiveness or quantity

oai_token-act-pair · gpt-3.5-turbo

the term "most" in various contexts indicating the greatest degree or quantity

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

rompt

-0.89

heid

-0.78

pload

-0.75

 Recording

-0.74

acus

-0.74

ategories

-0.72

amily

-0.69

ufact

-0.67

 Films

-0.67

cia

-0.65

POSITIVE LOGITS

 important

0.81

 interesting

0.80

 egregious

0.77

 likely

0.77

 bang

0.77

 expensive

0.76

 stringent

0.76

 favourable

0.76

 restrictive

0.76

 efficient

0.76

Activations Density 0.068%

superlatives related to effectiveness or quantity

the term "most" in various contexts indicating the greatest degree or quantity

No Comments

No Known Activations