INDEX

Explanations

phrases related to events or statements, potentially related to news or updates

oai_token-act-pair · gpt-3.5-turbo

the word "a" and its various contexts in document structures

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

alties

-0.93

Versions

-0.88

acers

-0.87

Ds

-0.83

rils

-0.81

byss

-0.79

bats

-0.77

ãĤ¨ãĥ«

-0.77

anism

-0.77

 stripes

-0.77

POSITIVE LOGITS

 continuation

1.03

 topic

1.00

 HUGE

0.98

 reminder

0.92

 serious

0.91

 tricky

0.91

 fascinating

0.91

 VERY

0.91

 conversation

0.89

guy

0.88

Activations Density 0.104%

phrases related to events or statements, potentially related to news or updates

the word "a" and its various contexts in document structures

No Comments

No Known Activations