INDEX

Explanations

references to modifications or alterations

oai_token-act-pair · gpt-3.5-turbo

the term "modified" in various contexts

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

ä

-0.79

çĦ

-0.77

arer

-0.76

HI

-0.76

ellow

-0.73

æµ

-0.71

Å

-0.71

 Glacier

-0.70

 Islanders

-0.69

Water

-0.69

POSITIVE LOGITS

 modifications

0.99

 modification

0.97

 carbohyd

0.90

 guiActiveUn

0.89

atile

0.88

imedia

0.83

 ende

0.83

 modifying

0.82

 nomine

0.82

 confir

0.81

Activations Density 0.013%

references to modifications or alterations

the term "modified" in various contexts

No Comments

No Known Activations