INDEX

Explanations

words related to expressing disagreements or differing opinions

oai_token-act-pair · gpt-3.5-turbo

instances of the word "how" in relation to discussions or debates

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

 Ferr

-0.69

 srfAttach

-0.66

isher

-0.63

Sao

-0.62

 Originally

-0.62

Sop

-0.61

Rib

-0.60

 Tanz

-0.60

 Examiner

-0.58

ãĥĩãĤ£

-0.58

POSITIVE LOGITS

soever

1.12

ever

0.73

 responsiveness

0.67

 THEY

0.67

 humans

0.67

beit

0.66

 mankind

0.64

binary

0.64

Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯

0.64

 exactly

0.63

Activations Density 0.120%

words related to expressing disagreements or differing opinions

instances of the word "how" in relation to discussions or debates

No Comments

No Known Activations