INDEX

Explanations

the word "seem" or its variations at various levels of intensity

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "seem" and its variations

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

estern

-0.99

ilts

-0.77

iding

-0.75

ests

-0.75

atching

-0.70

venge

-0.69

rench

-0.69

rouse

-0.69

orthern

-0.69

urrence

-0.66

POSITIVE LOGITS

rils

0.90

 awfully

0.79

 innocuous

0.78

ãĤ¨

0.78

 oddly

0.78

Pause

0.77

lessly

0.75

Magikarp

0.75

 strangely

0.75

 plaus

0.74

Activations Density 0.045%

the word "seem" or its variations at various levels of intensity

occurrences of the word "seem" and its variations

No Comments

No Known Activations