INDEX

Explanations

terms related to different types of objects or metaphorical concepts

oai_token-act-pair · gpt-3.5-turbo

metaphorical language related to animals and objects used to illustrate concepts or situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

AMI

-0.63

VICE

-0.63

 Dresden

-0.58

rity

-0.58

Ability

-0.58

 degraded

-0.58

NZ

-0.58

kson

-0.55

 Recre

-0.55

HCR

-0.55

POSITIVE LOGITS

bowl

0.79

 fide

0.78

headed

0.78

neck

0.76

bag

0.76

pit

0.73

tight

0.73

wagon

0.73

 wagon

0.69

cake

0.69

Activations Density 0.491%

terms related to different types of objects or metaphorical concepts

metaphorical language related to animals and objects used to illustrate concepts or situations

No Comments

No Known Activations