INDEX

Explanations

verbs followed by the word 'to'

oai_token-act-pair · gpt-3.5-turbo

clauses expressing likelihood or probability of events

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

DAQ

-0.80

soDeliveryDate

-0.73

Cooldown

-0.73

engineering

-0.70

illin

-0.69

hai

-0.68

anchester

-0.66

umption

-0.65

 Closing

-0.65

limits

-0.65

POSITIVE LOGITS

 engage

1.14

 participate

1.13

be

1.10

 receive

1.10

 absorb

1.03

 earn

1.03

 contribute

1.02

 incorporate

1.01

 enter

1.01

 accumulate

1.01

Activations Density 0.073%

verbs followed by the word 'to'

clauses expressing likelihood or probability of events

No Comments

No Known Activations