INDEX

Explanations

keywords related to impact or influence

oai_token-act-pair · gpt-3.5-turbo

mentions of the term "imp" and its variations in the context of awards and recognition

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

creen

-0.82

ãģ®éŃĶ

-0.78

çĦ

-0.77

 greens

-0.68

RAFT

-0.67

gaard

-0.67

TERN

-0.66

 UNIVERS

-0.66

Downloadha

-0.65

mberg

-0.65

POSITIVE LOGITS

acts

1.07

ressive

1.03

Imp

0.97

ulse

0.96

actor

0.96

acet

0.94

aired

0.91

ropri

0.90

ression

0.90

regn

0.89

Activations Density 0.008%

keywords related to impact or influence

mentions of the term "imp" and its variations in the context of awards and recognition

No Comments

No Known Activations