INDEX

Explanations

dates in the format of years

oai_token-act-pair · gpt-3.5-turbo

instances of the word "and" in various contexts

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

natureconservancy

-0.99

BIL

-0.83

ellectual

-0.71

ela

-0.69

sic

-0.69

Reason

-0.67

Ds

-0.67

é¾įå

-0.66

 Doodle

-0.66

uci

-0.66

POSITIVE LOGITS

 onwards

1.17

 thereafter

1.14

 onward

1.10

 again

1.07

1.06

1.04

 beyond

1.04

Activations Density 0.115%

dates in the format of years

instances of the word "and" in various contexts

No Comments

No Known Activations