INDEX

Explanations

specific words in the text

oai_token-act-pair · gpt-3.5-turbo

terms related to words and their correct usage

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

NetMessage

-0.89

âĹ¼

-0.79

aukee

-0.79

vic

-0.74

ramid

-0.71

ECA

-0.69

ummies

-0.68

outube

-0.67

feat

-0.66

erial

-0.66

POSITIVE LOGITS

 itself

1.05

ultimate

0.74

tymology

0.71

 spelled

0.68

scape

0.67

 Jesus

0.66

GOD

0.66

 interchange

0.66

PAT

0.66

Activations Density 0.087%

specific words in the text

terms related to words and their correct usage

No Comments

No Known Activations