INDEX

Explanations

mentions of specific names, likely proper nouns

oai_token-act-pair · gpt-3.5-turbo

names and terms related to specific individuals and styles, particularly in the context of architecture and pop culture

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

 concess

-0.79

icum

-0.74

 provision

-0.73

inent

-0.71

stru

-0.70

etheus

-0.70

ocrin

-0.69

sole

-0.68

 compr

-0.67

iden

-0.67

POSITIVE LOGITS

glers

1.15

vernment

0.96

 Pengu

0.84

irlfriend

0.81

ospels

0.80

IVERS

0.77

ORGE

0.77

ourmet

0.77

gets

0.76

FFER

0.75

Activations Density 0.059%

mentions of specific names, likely proper nouns

names and terms related to specific individuals and styles, particularly in the context of architecture and pop culture

No Comments

No Known Activations