INDEX

Explanations

terms related to technology and objects that can be manipulated or changed

oai_token-act-pair · gpt-3.5-turbo

nouns and specific terminology related to various subjects, including technology, healthcare, and environment

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

-0.69

Redd

-0.63

ZI

-0.63

berra

-0.61

 ..."

-0.60

Merit

-0.60

ãģ¾

-0.59

Enlarge

-0.57

Eva

-0.54

photos

-0.52

POSITIVE LOGITS

rul

0.73

 itself

0.68

 consists

0.68

 consisted

0.66

 fared

0.65

 cannot

0.65

 differs

0.60

 comprises

0.60

may

0.60

 behaves

0.59

Activations Density 0.966%

terms related to technology and objects that can be manipulated or changed

nouns and specific terminology related to various subjects, including technology, healthcare, and environment

No Comments

No Known Activations