INDEX

Explanations

words related to thoughts, expectations, or assumptions

oai_token-act-pair · gpt-3.5-turbo

expressions of conjecture or expectation

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

ffect

-0.70

 scrim

-0.70

prototype

-0.66

 sche

-0.63

CVE

-0.63

ammy

-0.63

 Funk

-0.61

versions

-0.60

 widow

-0.60

irst

-0.60

POSITIVE LOGITS

 Moreno

0.73

 ---------

0.69

 Trace

0.66

ername

0.66

ãĤ¦ãĤ¹

0.64

imaru

0.64

enance

0.64

rehens

0.63

uces

0.63

gat

0.63

Activations Density 0.125%

words related to thoughts, expectations, or assumptions

expressions of conjecture or expectation

No Comments

No Known Activations