INDEX

Explanations

mentions of the word "ens" followed by various numbers, potentially related to quantities or rankings

oai_token-act-pair · gpt-3.5-turbo

variations of the word "sense."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 hitch

-0.69

Tou

-0.63

 Panther

-0.62

 Barnett

-0.62

 edited

-0.61

holding

-0.61

 Hitman

-0.60

 Pell

-0.59

 penny

-0.59

 Breaker

-0.57

POSITIVE LOGITS

ens

1.20

ource

0.96

terday

0.94

 confir

0.90

ensions

0.87

ensibly

0.85

ailing

0.85

rir

0.85

burgh

0.84

ources

0.84

Activations Density 0.014%

mentions of the word "ens" followed by various numbers, potentially related to quantities or rankings

variations of the word "sense."

No Comments

No Known Activations