INDEX

Explanations

words related to geographical locations or unique identifiers

oai_token-act-pair · gpt-3.5-turbo

specific names and terms related to organizations, events, and entities, particularly in military and polling 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

etts

-0.94

hart

-0.87

nect

-0.85

ities

-0.80

cker

-0.79

quartered

-0.79

ggie

-0.78

ets

-0.76

aukee

-0.73

liest

-0.73

POSITIVE LOGITS

åĭ

0.83

enance

0.83

arget

0.79

ä»£

0.79

 Methods

0.77

 Yose

0.74

à¤

0.74

Sut

0.74

æĺ

0.72

het

0.71

Activations Density 0.028%

words related to geographical locations or unique identifiers

specific names and terms related to organizations, events, and entities, particularly in military and polling contexts

No Comments

No Known Activations