INDEX

Explanations

names or parts of names related to specific individuals or entities

oai_token-act-pair · gpt-3.5-turbo

words relating to political figures and their actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

nesday

-0.81

glers

-0.75

ancial

-0.71

rongh

-0.71

erential

-0.68

auga

-0.66

ankind

-0.65

awed

-0.64

inctions

-0.63

invoke

-0.63

POSITIVE LOGITS

ondo

0.67

tes

0.65

itect

0.64

 Odyssey

0.64

inian

0.63

 Cortex

0.62

ansas

0.61

ICT

0.60

 Lange

0.60

 Awakening

0.60

Activations Density 0.084%

names or parts of names related to specific individuals or entities

words relating to political figures and their actions

No Comments

No Known Activations