INDEX

Explanations

instances of words related to filtering or sifting

oai_token-act-pair · gpt-3.5-turbo

mentions of "filter" and related concepts

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

eanor

-0.81

nee

-0.74

pton

-0.68

erald

-0.68

NAS

-0.68

send

-0.66

peak

-0.65

ciating

-0.65

DAY

-0.65

psc

-0.63

POSITIVE LOGITS

 filters

1.22

 filter

1.17

 filtering

1.17

 Filter

1.10

filter

1.00

 filtered

0.93

Filter

0.83

 cutoff

0.82

 pores

0.78

 "$:/

0.77

Activations Density 0.010%

instances of words related to filtering or sifting

mentions of "filter" and related concepts

No Comments

No Known Activations