INDEX

Explanations

video identifiers from YouTube

oai_token-act-pair · gpt-3.5-turbo

variables and specific identifiers, likely related to programming or coding syntax

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

 Levant

-0.72

 coord

-0.71

 confir

-0.69

aneers

-0.63

Phones

-0.62

lined

-0.62

anwhile

-0.62

 estab

-0.62

anos

-0.58

 coma

-0.57

POSITIVE LOGITS

bi

0.81

xus

0.79

oda

0.71

chemy

0.70

Yu

0.67

rice

0.67

akespe

0.67

oa

0.66

witch

0.66

Activations Density 0.040%

video identifiers from YouTube

variables and specific identifiers, likely related to programming or coding syntax

No Comments

No Known Activations