INDEX

Explanations

news headlines including the abbreviation "AP"

oai_token-act-pair · gpt-3.5-turbo

occurrences of citations or references in news articles

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

angan

-0.62

course

-0.62

antha

-0.61

ts

-0.58

enges

-0.57

 scale

-0.57

Shape

-0.56

tons

-0.55

irection

-0.54

Scroll

-0.52

POSITIVE LOGITS

 CITY

0.74

PROV

0.70

 COUNTY

0.70

 STATE

0.67

COL

0.67

+++

0.66

—

0.64

 MARK

0.62

COR

0.62

COL

0.61

Activations Density 0.027%

news headlines including the abbreviation "AP"

occurrences of citations or references in news articles

No Comments

No Known Activations