INDEX

Explanations

mentions of a particular word with a common prefix 'Inf' followed by a varying number

oai_token-act-pair · gpt-3.5-turbo

occurrences of the term "Inf" in various contexts

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

\\\\\\\\

-0.96

âķĲâķĲ

-0.87

manship

-0.86

BOOK

-0.82

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.75

ãģį

-0.74

åĳ

-0.73

ãģ®éŃĶ

-0.72

å§«

-0.72

çĶ°

-0.71

POSITIVE LOGITS

rared

1.15

lamm

1.14

iltr

1.13

ertility

1.08

ractions

0.99

ocom

0.99

inia

0.95

ortun

0.95

requently

0.95

inite

0.95

Activations Density 0.009%

mentions of a particular word with a common prefix 'Inf' followed by a varying number

occurrences of the term "Inf" in various contexts

No Comments

No Known Activations