INDEX

Explanations

patterns of characters that don't form any coherent words or phrases

oai_token-act-pair · gpt-3.5-turbo

recurrent numerical values or statistics related to various topics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

 oath

-0.98

 prosec

-0.78

act

-0.78

 indoctr

-0.78

 condone

-0.77

 unsuspecting

-0.73

 freedom

-0.73

 ethos

-0.73

 initi

-0.72

 innocence

-0.72

POSITIVE LOGITS

Compared

1.13

Average

1.10

Still

1.05

Interest

1.04

Overall

1.04

Compar

1.04

Correct

1.00

Statistics

1.00

Compare

1.00

TABLE

1.00

Activations Density 0.529%

patterns of characters that don't form any coherent words or phrases

recurrent numerical values or statistics related to various topics

No Comments

No Known Activations