INDEX

Explanations

phrases related to specific frequencies and schedules

oai_token-act-pair · gpt-3.5-turbo

phrases related to numerical or time-based restrictions and conditions

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

Reviewer

-0.57

 Vlad

-0.54

 Abdul

-0.53

 Riyadh

-0.53

 Damascus

-0.52

Naj

-0.52

STER

-0.51

found

-0.51

 Trudeau

-0.51

NR

-0.50

POSITIVE LOGITS

å§«

0.71

secure

0.66

utenberg

0.65

 sqor

0.63

ploy

0.62

onym

0.59

aldo

0.58

rug

0.58

-)

0.58

daq

0.55

Activations Density 1.643%

phrases related to specific frequencies and schedules

phrases related to numerical or time-based restrictions and conditions

No Comments

No Known Activations