INDEX

Explanations

numerical values or codes within a structured dataset

oai_token-act-pair · gpt-3.5-turbo

numerical ratings and measurable data points

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

furt

-0.81

ierrez

-0.73

abwe

-0.69

 disadvant

-0.66

isphere

-0.65

etsk

-0.65

 tightening

-0.64

undai

-0.63

 womb

-0.62

 infl

-0.62

POSITIVE LOGITS

HOU

0.74

px

0.74

$$$$

0.72

pt

0.71

BR

0.69

 Duration

0.68

0.66

 Comments

0.65

Activations Density 0.226%

numerical values or codes within a structured dataset

numerical ratings and measurable data points

No Comments

No Known Activations