INDEX

Explanations

verbs related to emphasizing or highlighting

oai_token-act-pair · gpt-3.5-turbo

terms related to affirmation and emphasis

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

 inaction

-0.73

 ACTION

-0.69

 folly

-0.66

 indifference

-0.65

 captive

-0.64

Reviewer

-0.63

 amusement

-0.62

 sterling

-0.61

 distraction

-0.61

 rationality

-0.60

POSITIVE LOGITS

istered

1.16

ited

1.13

ighed

1.12

uates

1.11

pling

1.10

uld

1.06

izes

1.05

tained

1.05

pped

1.05

pled

1.05

Activations Density 0.184%

verbs related to emphasizing or highlighting

terms related to affirmation and emphasis

No Comments

No Known Activations