INDEX

Explanations

phrases related to hierarchical classification or ranking

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

hur

-0.76

okia

-0.71

ãĤŃ

-0.68

mercial

-0.66

alam

-0.65

ãĤ¸

-0.64

issance

-0.64

Frames

-0.63

 pron

-0.62

aston

-0.61

POSITIVE LOGITS

 tiers

0.91

 tier

0.87

uple

0.85

neys

0.81

rait

0.81

arity

0.75

ney

0.75

raits

0.74

 VIII

0.72

lists

0.71

Activations Density 0.026%

phrases related to hierarchical classification or ranking

No Comments

No Known Activations