INDEX

Explanations

mentions or references to specific names, likely related to news or media coverage

oai_token-act-pair · gpt-3.5-turbo

names of individuals and locations

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

ISA

-0.81

SPONSORED

-0.78

SPA

-0.73

acebook

-0.71

 srfAttach

-0.66

Applications

-0.64

SAY

-0.63

MAP

-0.61

Ü

-0.60

leness

-0.60

POSITIVE LOGITS

vous

1.05

ÅĤ

0.76

hur

0.73

Ã¡n

0.72

ovsky

0.70

Ö¼

0.68

henko

0.67

ooth

0.67

owicz

0.67

opol

0.66

Activations Density 0.302%

mentions or references to specific names, likely related to news or media coverage

names of individuals and locations

No Comments

No Known Activations