INDEX

Explanations

references to various adventurous activities or contexts, particularly related to levels of adventure

oai_token-act-pair · gpt-3.5-turbo

references to the concept of "adventure" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 obligated

-0.82

heed

-0.78

early

-0.72

ombo

-0.71

 dotted

-0.71

 iodine

-0.69

iod

-0.69

animous

-0.68

 temporal

-0.68

 veget

-0.67

POSITIVE LOGITS

 Adventure

1.09

Adventure

0.97

nels

0.85

glers

0.83

 Railroad

0.81

rs

0.78

schild

0.78

cade

0.77

 Mansion

0.77

 Ashe

0.77

Activations Density 0.009%

references to various adventurous activities or contexts, particularly related to levels of adventure

references to the concept of "adventure" in various contexts

No Comments

No Known Activations