INDEX

Explanations

words related to trends or patterns

oai_token-act-pair · gpt-3.5-turbo

references to trends or patterns

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

gur

-0.98

tein

-0.74

 oath

-0.68

Lak

-0.65

hiro

-0.64

 lungs

-0.63

idden

-0.63

 barr

-0.61

oÄŁ

-0.61

 Aires

-0.60

POSITIVE LOGITS

setting

1.20

 trend

1.05

 Trend

1.03

line

0.91

Trend

0.90

 trends

0.89

icity

0.83

set

0.83

 Trends

0.82

lines

0.81

Activations Density 0.015%

words related to trends or patterns

references to trends or patterns

No Comments

No Known Activations