INDEX

Explanations

words related to events or activities that involve testing or demonstrating something

oai_token-act-pair · gpt-3.5-turbo

variations of the word "test."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 channel

-0.62

man

-0.61

 absent

-0.61

 copy

-0.60

 disappeared

-0.60

 painter

-0.59

 articulate

-0.58

 mortg

-0.57

 position

-0.56

tone

-0.56

POSITIVE LOGITS

ests

4.69

esting

2.38

ested

2.21

est

2.08

EST

2.05

esters

1.81

ester

1.48

estation

1.38

irement

1.29

esta

1.27

Activations Density 0.006%

words related to events or activities that involve testing or demonstrating something

variations of the word "test."

No Comments

No Known Activations