INDEX

Explanations

terms related to mass incidents or actions

oai_token-act-pair · gpt-3.5-turbo

terms related to mass events, particularly mass casualties or incidents

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

ĪĴ

-0.81

worldly

-0.65

NZ

-0.64

Desk

-0.61

peg

-0.59

 Ital

-0.59

#$#$

-0.59

ĵĺ

-0.59

 Aires

-0.58

MAC

-0.58

POSITIVE LOGITS

ãĤ¼ãĤ¦ãĤ¹

0.83

endars

0.75

usions

0.70

uity

0.69

luster

0.67

achusetts

0.67

atoes

0.66

utation

0.65

 Manson

0.65

abulary

0.65

Activations Density 0.125%

terms related to mass incidents or actions

terms related to mass events, particularly mass casualties or incidents

No Comments

No Known Activations