INDEX

Explanations

numbers and words related to comparisons

oai_token-act-pair · gpt-3.5-turbo

references to significant discussions or occurrences

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

 passports

-0.70

 coli

-0.69

 Peaks

-0.66

 throats

-0.65

 DISTRICT

-0.64

 Rover

-0.59

 beams

-0.59

 assassins

-0.57

CHAT

-0.57

 Advertisement

-0.56

POSITIVE LOGITS

WER

0.82

rah

0.69

ãĥ´ãĤ¡

0.67

needed

0.67

uable

0.66

Ïģ

0.63

ashtra

0.62

erous

0.60

cape

0.60

enos

0.60

Activations Density 0.278%

numbers and words related to comparisons

references to significant discussions or occurrences

No Comments

No Known Activations