INDEX

Explanations

mentions of the word "flagship."

oai_token-act-pair · gpt-3.5-turbo

the term "flagship" in relation to products or services

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Rule

-0.62

Rule

-0.61

Abd

-0.60

cook

-0.60

 Tolkien

-0.60

ould

-0.60

 Chomsky

-0.60

Rep

-0.59

Eck

-0.58

udic

-0.58

POSITIVE LOGITS

 flagship

3.61

 centerpiece

1.16

 marquee

1.14

 cornerstone

1.11

 cruiser

1.11

 Cruiser

1.09

 Flags

1.09

flags

1.07

 powerhouse

1.06

 mascot

1.05

Activations Density 0.016%

mentions of the word "flagship."

the term "flagship" in relation to products or services

No Comments

No Known Activations