INDEX

Explanations

time-related information, specifically total time measurements

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "Total" and related phrases indicating summaries of quantities or counts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

etsy

-0.86

pher

-0.83

 Lazarus

-0.79

*/(

-0.70

pload

-0.70

 pedd

-0.69

acle

-0.68

 promul

-0.67

bub

-0.67

lyn

-0.65

POSITIVE LOGITS

 Total

0.84

itarian

0.83

ROM

0.75

Pros

0.72

NL

0.72

 Estimated

0.71

 Conversion

0.71

CPU

0.71

Total

0.69

 Recall

0.69

Activations Density 0.009%

time-related information, specifically total time measurements

occurrences of the word "Total" and related phrases indicating summaries of quantities or counts

No Comments

No Known Activations