INDEX

Explanations

instances of the word "underlying" or phrases related to underlying factors or structures

oai_token-act-pair · gpt-3.5-turbo

references to underlying concepts or factors

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

uden

-0.81

cture

-0.80

hops

-0.79

aterasu

-0.78

ischer

-0.77

UNCH

-0.76

uncle

-0.76

asia

-0.75

isters

-0.75

ishers

-0.75

POSITIVE LOGITS

 assumptions

1.00

 assumption

0.99

 fundamentals

0.97

 principles

0.97

 premise

0.97

 structure

0.96

 motivations

0.96

 structures

0.93

 infrastructure

0.93

 theme

0.90

Activations Density 0.035%

instances of the word "underlying" or phrases related to underlying factors or structures

references to underlying concepts or factors

No Comments

No Known Activations