INDEX

Explanations

instances where something is being limited, furthered, or only allowed to a certain extent

oai_token-act-pair · gpt-3.5-turbo

words that indicate limitations or constraints

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

)|

-0.64

âĿ

-0.61

},{"

-0.61

},"

-0.60

ilege

-0.60

oir

-0.60

 Base

-0.59

oko

-0.57

onomy

-0.57

 Hold

-0.56

POSITIVE LOGITS

 preferring

1.15

 suggesting

1.05

 implying

0.97

 noting

0.97

 adding

0.93

 culminating

0.92

 spilling

0.90

 emphasizing

0.89

 prompting

0.89

 echoing

0.86

Activations Density 0.421%

instances where something is being limited, furthered, or only allowed to a certain extent

words that indicate limitations or constraints

No Comments

No Known Activations