INDEX

Explanations

phrases related to specific people or entities

oai_token-act-pair · gpt-3.5-turbo

references to specific types or names, particularly related to "obi" and potential topics like characters or entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

shall

-0.77

âķĲâķĲ

-0.72

ELL

-0.66

PID

-0.65

luck

-0.64

lling

-0.64

DMV

-0.63

liam

-0.63

rab

-0.63

rier

-0.62

POSITIVE LOGITS

obi

1.42

oses

1.12

Wan

1.10

OGR

1.02

ograp

0.99

ographical

0.98

ography

0.97

ographer

0.96

ographies

0.96

oche

0.96

Activations Density 0.009%

phrases related to specific people or entities

references to specific types or names, particularly related to "obi" and potential topics like characters or entities

No Comments

No Known Activations