INDEX

Explanations

elements of computer code or web links

oai_token-act-pair · gpt-3.5-turbo

patterns of alphanumeric sequences or identifiers

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 Franch

-0.67

 Prim

-0.65

 monop

-0.63

Rug

-0.63

HEL

-0.62

Opt

-0.62

 Polar

-0.58

Hok

-0.58

 Subst

-0.58

Dol

-0.57

POSITIVE LOGITS

ertodd

0.77

"],"

0.72

Ñı

0.71

¿½

0.71

antically

0.69

WER

0.68

 "$:/

0.68

]+

0.68

Ĥ¬

0.68

ilee

0.66

Activations Density 0.093%

elements of computer code or web links

patterns of alphanumeric sequences or identifiers

No Comments

No Known Activations