INDEX

Explanations

mentions of specific names or terms related to various contexts which include individuals, locations, and concepts

oai_token-act-pair · gpt-3.5-turbo

occurrences of names and references related to particular individuals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 elig

-0.94

 bluff

-0.69

lished

-0.62

 diploma

-0.60

 entrants

-0.60

 runway

-0.57

$$$$

-0.57

 seism

-0.57

 crawl

-0.56

RU

-0.56

POSITIVE LOGITS

igans

0.96

felt

0.84

frames

0.76

oused

0.73

chenko

0.73

nian

0.73

velt

0.71

aku

0.71

zhen

0.68

idan

0.67

Activations Density 0.108%

mentions of specific names or terms related to various contexts which include individuals, locations, and concepts

occurrences of names and references related to particular individuals

No Comments

No Known Activations