INDEX

Explanations

references to authoritative figures or entities, such as leaders, officials, or iconic characters

oai_token-act-pair · gpt-3.5-turbo

instances of the word "arch" 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 @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Kamp

-0.75

uana

-0.71

 Pearce

-0.68

ABE

-0.67

ggle

-0.66

leneck

-0.65

ORTS

-0.63

TPS

-0.63

KK

-0.62

Ô

-0.61

POSITIVE LOGITS

ipel

1.35

itect

1.15

bishop

1.11

adia

0.98

ival

0.98

ivist

0.94

di

0.93

aic

0.93

esian

0.85

rival

0.85

Activations Density 0.023%

references to authoritative figures or entities, such as leaders, officials, or iconic characters

instances of the word "arch" in various contexts

No Comments

No Known Activations