INDEX

Explanations

verbs or phrases relating to analysis, definition, or description

oai_token-act-pair · gpt-3.5-turbo

action-oriented verbs and descriptors related to processes or states

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

Jah

-0.67

along

-0.66

amen

-0.63

HQ

-0.63

hillary

-0.62

asi

-0.60

Stream

-0.58

igers

-0.58

CHAT

-0.57

atro

-0.57

POSITIVE LOGITS

Initialized

0.82

isable

0.76

iated

0.70

ifiable

0.68

aneously

0.66

 è£ıç

0.66

 thous

0.66

abal

0.64

(<

0.63

 dracon

0.63

Activations Density 0.659%

verbs or phrases relating to analysis, definition, or description

action-oriented verbs and descriptors related to processes or states

No Comments

No Known Activations