INDEX

Explanations

phrases indicating the speaker's personal opinions or analysis

oai_token-act-pair · gpt-3.5-turbo

phrases expressing personal opinions or perceptions

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

ories

-0.68

 Split

-0.65

Split

-0.62

iability

-0.62

 McKin

-0.61

oresc

-0.61

BRE

-0.59

atters

-0.59

 Paste

-0.58

Tra

-0.58

POSITIVE LOGITS

 detriment

0.90

 converter

0.81

 liking

0.76

 extent

0.72

selves

0.72

venge

0.71

merce

0.70

 mailbox

0.69

©¶æ

0.65

 hemisphere

0.65

Activations Density 0.087%

phrases indicating the speaker's personal opinions or analysis

phrases expressing personal opinions or perceptions

No Comments

No Known Activations