INDEX

Explanations

references to news articles or media

oai_token-act-pair · gpt-3.5-turbo

references to news and media-related terms

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

inho

-0.70

Lak

-0.64

 Stark

-0.63

 Wade

-0.63

asse

-0.63

 Defin

-0.63

 Alto

-0.62

 Greenwood

-0.61

 Scher

-0.61

zx

-0.61

POSITIVE LOGITS

room

1.23

rooms

1.23

worthy

1.22

feed

1.20

worthiness

1.15

reader

1.11

agents

1.05

weekly

1.04

agency

1.01

peak

1.01

Activations Density 0.053%

references to news articles or media

references to news and media-related terms

No Comments

No Known Activations