INDEX

Explanations

terms related to progress, change, and advancement

oai_token-act-pair · gpt-3.5-turbo

terms associated with improvement and significant changes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 beans

-0.65

rich

-0.64

 Beans

-0.64

 compuls

-0.63

rive

-0.62

milo

-0.62

amina

-0.59

 bean

-0.57

 dinners

-0.55

ictions

-0.55

POSITIVE LOGITS

 compared

0.92

 departure

0.80

 leaps

0.68

ahead

0.67

 considering

0.67

atform

0.66

kees

0.66

ulation

0.65

thal

0.65

 offensively

0.65

Activations Density 0.119%

terms related to progress, change, and advancement

terms associated with improvement and significant changes

No Comments

No Known Activations