INDEX

Explanations

references to educational grades, particularly focusing on the grade 9 level

oai_token-act-pair · gpt-3.5-turbo

references to grading levels in educational contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Bundy

-0.98

riks

-0.78

ĸļ

-0.75

orial

-0.72

pload

-0.72

othe

-0.72

 Liberties

-0.71

ople

-0.71

sle

-0.71

gren

-0.70

POSITIVE LOGITS

 grade

0.93

 grades

0.89

 rating

0.89

Grade

0.83

 plutonium

0.82

 grading

0.77

ably

0.76

 Grade

0.73

 ratings

0.68

 Rated

0.67

Activations Density 0.013%

references to educational grades, particularly focusing on the grade 9 level

references to grading levels in educational contexts

No Comments

No Known Activations