INDEX

Explanations

words related to the concept of "saving"

oai_token-act-pair · gpt-3.5-turbo

references to specific individuals or groups within narratives

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 trademark

-0.71

 resolve

-0.69

 unfor

-0.67

 inclined

-0.67

 predis

-0.66

 Turing

-0.65

backer

-0.65

 phenomen

-0.65

 pointer

-0.64

 hypert

-0.64

POSITIVE LOGITS

irus

1.26

ille

1.14

arro

1.12

atars

1.11

oided

1.08

atar

1.06

ihara

1.02

olt

1.01

illa

1.00

ocado

1.00

Activations Density 0.011%

words related to the concept of "saving"

references to specific individuals or groups within narratives

No Comments

No Known Activations