INDEX

Explanations

phrases related to tasks or duties given to someone

oai_token-act-pair · gpt-3.5-turbo

references to assignments

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Habit

-0.75

ASE

-0.67

ells

-0.67

GI

-0.65

mph

-0.65

Ïī

-0.64

twitter

-0.63

Som

-0.62

 supra

-0.62

 Horton

-0.60

POSITIVE LOGITS

 assignment

1.43

 assignments

1.23

 Assignment

1.04

 ende

0.84

 assigned

0.83

eering

0.80

 allocation

0.78

ignment

0.77

 assign

0.76

idon

0.76

Activations Density 0.007%

phrases related to tasks or duties given to someone

references to assignments

No Comments

No Known Activations